摘要
增强SAR舰船尾迹图像中模糊的开尔文尾迹并保持湍流尾迹特征对舰船及运动参数的反演具有重要作用。该文利用快速自适应2维经验模式分解方法(Fast and Adaptive Bidimensional Empirical Mode Decomposition,FABEMD)实现图像中开尔文尾迹,湍流尾迹和其他中/大尺度海洋特征的分解,提高开尔文尾迹相对其他特征的图像和频谱对比度。同时引入并改进干涉图Goldstein滤波器实现对开尔文尾迹的进一步增强,并利用不变矩对增强后的SAR舰船尾迹图像进行评价。通过原理分析、增强实验和主/客观评价,表明该方法具有显著的开尔文尾迹增强效果,并保持了湍流尾迹特征,实现效率高且适用性较强。
Enhanced SAR ship wake images with blur Kelvin wakes and reserved turbulent wakes are very important to the inversions of ship and motion parameters. This paper applies the Fast and Adaptive Bidimensional Empirical Mode Decomposition （FABEMD） to decompose the SAR ship wake image into Kelvin wakes, turbulent wakes and other multiscale features, which enhances the gray intensity and spectrum contrast of Kelvin wakes to other features. Based on the FABEMD, a modified Goldstein interferogram filter is developed to further enhance the Kelvin wakes. Moreover, the moment invariants are introduced to evaluate the enhancement. Therefore, the Kelvin wakes are dramatically enhanced and the turbulent wakes are reserved. Algorithm analysis, experiments, subjective and objective evaluations show the reasonable efficiency and capabilities.
基金
国家部委预研基金资助课题
关键词
合成孔径雷达
图像增强
尾迹
2维经验模式分解
快速自适应2维经验模式分解
干涉相位滤波
不变矩
Synthetic Aperture Radar （SAR）
Image enhancement
Ship wake
Bidimensional Empirical Mode Decomposition （BEMD）
Fast and Adaptive Bidimensional Empirical Mode Decomposition （FABEMD）
Interferogram filter
Moment invariants
